The Three Walls
What the press translation missed in Chey Tae-won’s National Assembly speech
Saint Clair · Market Intelligence | May 2026
Korea’s most senior business voice articulated a doctrine on three dimensions of national confinement. The wire-service summary covered roughly a third of it. Allocators should read the full text.
Reading the Whole Speech
By the morning after Chey Tae-won spoke at the National Assembly on 28 April, the wire-service summary read: he proposed economic integration with Japan. The summary was accurate, and it covered roughly a third of what he said. Read in full, the speech is a doctrine on three dimensions of Korean confinement, with one through-line beneath: at national scale, Korea cannot resolve any of them.
Three institutional facts shape the speech. Chey chairs SK Group. He also chairs the Korea Chamber of Commerce and Industry, through which he sustains a regular policy dialogue with Keidanren. Separately, he has run an eleven-year research line on the measurement of social value, with its own institute and tested methodology. The two chairmanships and the research line together give this speech the institutional weight that justifies reading it in full.
In his own framing, the speech sets out three problems. The first is capitalism: a slowing growth model now facing an AI shock measured in decades. The second is AI infrastructure: at the scale the next decade requires, Korea cannot mobilise the capital alone. The third is geopolitics: at one-tenth of China’s GDP, Korea cannot set the rules under which any of this gets built.
The single line uniting them runs beneath the surface. In each leg, Chey proposes the construction of a frame larger than the national one. The walls follow.
A Measurement Problem
Begin with capitalism’s own diagnosis. The dual engine of Korea’s post-war development (economic growth and democratisation) worked, and now slows on the growth side. Capital efficiency is falling. Returns from investment are thinning. The development model that produced the corporate economy now needs rewriting before the AI shock arrives, because the shock will run for a decade and reshape labour structures faster than retraining can absorb them. Existing instruments were calibrated for ordinary structural change. The velocity of AI substitution exceeds them.
His diagnosis is that capitalism mismeasures what it produces. The system counts monetary output. The jobs companies create, the environmental cost they avoid, the welfare contribution they make, the tax they pay — all are real outputs, all unpriced by markets, while governments collect taxes on monetary output and provide welfare downstream. The gap between what the economy produces and what the economy measures is, in Chey’s account, the reason public policy cannot keep pace with the AI shock.
This frame runs in direct parallel to the Beyond GDP debate alive in international scholarship since the Stiglitz–Sen–Fitoussi Commission’s 2009 report on economic performance and social progress. Diane Coyle’s GDP: A Brief but Affectionate History (2014) traced the conceptual ceiling of monetary aggregates as performance measures. Mariana Mazzucato’s The Value of Everything (2018) extended the critique into who creates value and who captures it. Chey has spent eleven years working the same problem from inside Korean industrial scale, with a research institute and tested methodology behind him. The leg situates Korea inside an ongoing international scholarly conversation. Western allocators who read it as Chey’s private programme will misread its weight.
The policy form he proposes is two markets running in parallel: one priced in money, one priced in social value, with the state acting as referee. The labour-absorption capacity of the second market grows as the first contracts under AI substitution.
Saint Clair engages this leg on the terms Chey set. The measurement argument has its own technical literature, its own scepticism, its own difficulty. What it represents in this speech is a frame: the Korean economy will need a different ledger when AI arrives at full force. The capital-allocation implication sits in the lead time. Policy will be built around a measurement architecture Korean allocators will see assembled before international ones do, and the year in which the asymmetry closes is the year the discount disappears.
Capital and Concrete
The infrastructure leg is denominated in numbers. Chey identifies four bottlenecks in the AI build-out. Capital, electricity, GPUs, and memory. Each constrains the others.
The capital number he uses is direct. A one-gigawatt AI data centre costs roughly fifty billion dollars to build. Korea’s total data-centre capacity stands at around one gigawatt; Chey puts the share that is genuinely AI-capable at under five per cent. He estimates the requirement, for Korea to participate as a builder of the infrastructure decade, at ten to thirty gigawatts. The arithmetic at the upper end is half a trillion dollars to a trillion-five.
Electricity has shifted from a utility input to a structural constraint, and the deployment model has changed with it. Trump’s instruction to the American hyperscalers, in Chey’s recounting, was to generate their own power. The new pattern co-locates AI data centres with dedicated power plants, decoupled from the central grid, because the grid cannot absorb AI load fluctuation efficiently. China has built energy generation capacity faster than anyone else and continues to widen that lead. The United States retains its lead on advanced chips. Korea sits in the memory chokepoint, where three companies globally (two of them Korean) supply the high-bandwidth memory that AI accelerators require.
The GPU position is fluid. Nvidia’s lead remains, but the inference market is segmenting and the dominance of any single architecture loosens as workload diversifies. The memory position is tighter and likely to remain so.
Read carefully, the leg makes one argument. The capital required for Korea to be a builder in the AI era exceeds what national mobilisation can supply. The grid logic the build-out requires exceeds what national grid planning can deliver. The supplier concentration on memory exceeds what an independent national strategy can hedge. Each, on Chey’s reading, is a wall a country of fifty-one million cannot scale alone. The capital-allocation question for the next decade is on what terms Korea participates and at what scale of partner the participation requires.
An Island That Needs to Stop Being One
The third wall is the leg the press took. Chey’s framing is geographic. Korea sits on a peninsula; the continent is sealed by DPRK; the southern coastline opens to ocean alone. Korea’s economy, in functional terms, is an island economy. Its GDP is roughly one-tenth of China’s. The asymmetry mirrors how Korea regards any country at one-tenth of its own size. The relationship runs one way.
His proposal is integration with Japan, in a form serious enough that the rest of the world reads the combined economy as a single market. Combined GDP at six trillion dollars sits at roughly a third of China’s. At that scale, Korea–Japan becomes a structure the United States and China can no longer route around. Below that scale, they will. Chey invokes the EU as the analogue, with the operative observation that the EU began with France and Germany; the twenty-seven came later, because twenty-seven cannot rule with speed and a small core can.
From the integrated core, two extensions follow. The southern: ASEAN economies gravitate toward a six-trillion-dollar anchor, and an Asian-style community begins to form around it. The northern: China’s coastal northeast (산동성, 강소성, the three north-eastern provinces) and Russia’s Primorsky enter the bloc’s economic orbit, and the resulting gravitational pressure expresses itself against DPRK closure. Both extensions are long-dated. Chey treats them as the long-term premium that follows from integration; integration must come first.
Concrete first steps appear in the speech and in the questions that followed. A thirty-item integration list. Cross-border power lines and gas interconnects with Japan. Joint energy-cost reduction through shared infrastructure. KCCI’s standing dialogue with the Japan Chamber of Commerce. Tokyo has not reciprocated publicly at the level of the proposal. Korean political opposition retains historical sensitivities any administration must price.
The thesis is doctrine running ahead of mechanism. The gap is real, and it is the reason the speech matters as a signal. Doctrine articulated with this institutional weight, in this venue, is the form by which mechanism eventually arrives. The capital-allocation implication is that the integration thesis will compound across cycles even as discrete bilateral mechanisms remain slow.
What the Allocator Does Differently
Three walls, one logic. Capital allocation is implicated at each layer, and at each layer the constraint sits above what national mobilisation can resolve.
The signal weight sits in the seniority of the voice and in the public record building behind it. The chairman of SK and of KCCI is the most senior business voice in Korea, and he chose the National Assembly to articulate this, alongside February’s trilateral cooperation push and an eleven-year research programme on social value measurement.
The allocator who has read the speech in full asks different questions of Korean opportunities. On the measurement leg: when the new architecture comes into Korean policy, and how long the allocation lead time runs before international markets price it in. On the infrastructure leg: at what gigawatt commitment Korea participates in the AI build-out, and with which counterparties. On the integration leg: how fast Chey’s thesis converts from doctrine into bilateral mechanism, and which sectors compound the premium first.
The discipline of reading what the establishment reads, in its full register, comes before any position worth taking. Allocators who skip it work from the wire-service summary.
The measurement and isolation walls have particular instantiations in Korea’s venture maturity profile and contract architecture, which the Ground Truth: Korea Corridor sub-series addresses.
Sources
Primary: Chey Tae-won, transcript of address to National Assembly seminar hosted by the Korea–China Parliamentarians’ Union, 28 April 2026 (provided by Saint Clair).
SK chairman calls for deeper South Korea-Japan economic ties — UPI, 28 Apr 2026
KCCI chief suggests EU-style economic bloc between Korea, Japan — Korea Herald
SK chief doubles down on Korea-Japan economic bloc push — Korea Herald
Chey Tae-won urges Korea-Japan economic bloc — AJU Press, 28 Apr 2026
Korea faces mounting calls for economic coalition with Japan — Korea Times, 15 Aug 2025
Stiglitz, J., Sen, A., & Fitoussi, J.-P. (2009). Report by the Commission on the Measurement of Economic Performance and Social Progress.
Coyle, D. (2014). GDP: A Brief but Affectionate History. Princeton University Press.
Mazzucato, M. (2018). The Value of Everything: Making and Taking in the Global Economy. Allen Lane.
Disclaimer: This article is for informational purposes only and does not constitute investment advice. All decisions should be made based on independent research and consultation with qualified advisors.
About Saint Clair — Advisory & Capital: Saint Clair designs and builds cross-border capital infrastructure between Europe and Asia — proposing access where access is scarce, and creating structure where structure is absent. Since 2016.
Learn more: saintclair.sg | Contact: contact@saintclair.sg